fleet-server performance improvements

In order to better assist scaling efforts we should audit the fleet-server codebase for any performance improvements we can make.

The performance improvements we should seek are:
- Reduce calls to elasticsearch where possible (merging two calls with a painless script is preferred over two separate ES updates)
- Reduce allocations in order to reduce heap size and reduce the work that the garbage collector has to do.
- Ensure struct alignment (easy to add automation to ensure this works)

Acceptance criteria:
- All changes must have benchmarks associated with them.
- Benchmarks should show a reduction in one of the tracked catagories
- Benchmark comparisons between the baseline and improvements - with `-memprofile` must be available for analysis

Focus will be directed towards high-use packages first.

No files, tests, or entry points are named; begin by identifying the high-use fleet-server packages and establishing baseline benchmarks. Done means each performance change has benchmark comparisons, including -memprofile data, and demonstrates reduced Elasticsearch calls, allocations, or struct-padding waste.

Written by the indexing model from the issue text.